Bourse
ENPL
Get the App
Google PlayAndroid appWindowsDownload the .exe installerWeb AppNo install needed
Menu
Data
Conviction ScoreEvery insider buy scored 0-100 by machine learningCoincidence TrackerCongress trades next to federal contractsEuropean InsidersDirector dealings: Warsaw, Frankfurt, Paris, LondonInsider TradesSEC Form 4 buys and sells by CEOs and directorsCongress TradesStock trades by Senators and RepresentativesCongress NetworkWhich politicians trade the same stocksTrump TradesThe President’s 278-T reports and portfolioCabinet & Fed TradesSecretaries, Fed governors, White House staffHedge Funds (13F)Buffett, Burry, Ackman and every big fundJim Cramer & CNBCEvery Mad Money call, scoredLobbying & Revolving DoorWho pays Washington, and who they hireGoogle Trends & WikipediaWhen the crowd suddenly caresWARN LayoffsMass layoff notices from 48 statesAll Datasets44 datasets: SEC, macro, sentiment
Strategies
AI StrategiesModels retrained monthly on every signalInverse CramerBet against Mad Money, or with itTrump PortfolioCopy the President’s stock buysCongress Long-ShortLong what Congress buys, short what it sellsCopy Buffett & BurryFamous 13F portfolios, rebalancedUpcoming TradesSee the next rebalance before it happens
Radar
Radar
API
DevelopersREST, Parquet, webhooks, Python, pricingAPI ReferenceEvery endpoint, field and webhook (docs.bourse.pl)Status and SLAMeasured uptime, 90 days (status.bourse.pl)API ProCommercial license, 250,000 requests a dayBourse Agent (AI)Send a filing, get a decision as JSONIntegrationsAWS, MCP Registry, GitHub, Discord, TelegramGet a Free Key1,000 requests a day, no cardOpenAPI SpecImport into Postman or a code generatorMCP for Claude & ChatGPTAsk your AI assistant about any filingWebhooksNew filings pushed to n8n, Zapier or your serverPython: bourse-quantpip install bourse-quant, pandas DataFrames
Learn
TutorialsGuides for the app, strategies and the APIBlogPublic filings, explainedMethodologyWhere every number comes fromPartner Program30% of every payment for 12 months
Pricing
Pricing
Developers / API
Developers / API
Get the App
Google PlayAndroid appWindowsDownload the .exe installerWeb AppNo install needed

Not every insider buy
is a signal.

The Bourse Conviction Score rates every open-market insider purchase from 0 to 100 with a machine-learning model, and shows the reasons: first buy in a long time, other executives buying the same week, a price near the 52-week low, a stake that doubled.

This purchase scores 100/100 on the Bourse Conviction Score, the highest for KOD in Bourse data (since June 4, 2026). (KOD, 2026-09-28)

See today’s scores GET /v1/conviction
100/ 100

Tested on months it never saw.

0.6114ROC AUC on unseen months
2026-08-25 – 2026-10-01, 1875 purchases
60%of the top-scored 10% landed in the best 30% of their month
base rate 28%
+7.8%median 1-month return of the top 10%
all purchases -3.0%
9,880Form 4 purchases in the training data
2026-05-11 – 2026-11-10

Highest scores, last 60 days.

ScoreTickerBuyerValueDateWhy
100KODInsider$20.2M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$17.7M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$10.4M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$9.6M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$9.3M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$8.7M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$7.4M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$6.6M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$5.8M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$5.7M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$5.5M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$4.7M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$3.9M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$3.7M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before
100KODInsider$3.3M2026-09-28Price 500% above the 52-week low · Shares rose 130% in the month before

Purchases filed more than 7 days ago. Live scores are in the app and the API.

Questions.

What does a score of 98 mean?

The model rated the purchase higher than 98% of insider purchases it was not trained on. It is a rank, not a probability of profit.

Which model is it?

Gradient-boosted decision trees with the XGBoost algorithm (logistic loss, second-order gain, L2 regularisation, learned direction for missing values), trained on Form 4 open-market purchases. The label is whether the stock’s 1-month return after the trade was in the top 30% of all insider purchases that month.

How do I get the score?

In the app (More, Conviction Score), in the API (GET /v1/conviction and the conviction_score field of /v1/insiders) and with pip install bourse-quant: bourse.get_conviction(min_score=90).

What are the limits?

The training history is short (Bourse Form 4 data since May 2026) and covers one market regime. Price features exist only where daily prices were available. The out-of-sample result is modest and can change. Treat the score as a way to sort filings, not as advice.

Bourse provides financial data and editorial information, parts of which are generated by AI. This page is for general education only. It is not personalised investment advice or a recommendation to buy or sell any financial instrument. Make your own decisions.